Cross-Matching in Liver Transplantation

Writing in this Journal in 1983, Alan Ting (1) suggested that the cross-match test may be “the most important contribution of the tissue typing laboratory to clinical renal transplantation”. The dire consequences of transplanting a kidney into a recipient with a positive cross-match are well known; however, liver transplants are routinely performed across this barrier without suffering the penalty of hyperacute rejection. Between 7% and 33% of patients awaiting their first liver transplant harbor cytotoxic HLA-specific antibodies, and at least half of these patients will be IgG anti-donor positive (Tables 1 and 2). It might be expected, therefore, that a large percentage of liver grafts would be lost to hyperacute rejection. In fact, only a few instances of hyperacute rejection have been reported (as discussed below). A positive donor cross-match at transplantation may also be associated with long-term effects on solid organ grafts. In heart transplantation, hyperacute rejection is rare, but transplantation against the cross-match barrier is associated with an increased risk of atherosclerosis (2) and lower graft survival (3-5). In this Overview, we consider the current evidence that the cross-match test may be of use in predicting outcome after liver transplantation and discuss some of the mechanisms that may enable liver transplants to resist both the immediate and long-term effects of alloantibody. HYPERACUTE REJECTION Animal models. Hyperacute rejection does occur in animal models where liver transplants are performed against positive cross-matches. Early experiments with rabbit-to-dog liver xenografts demonstrated consumption of immunoglobulin, platelets, leukocytes, complement, and clotting factors, phenomena associated with hyperacute rejection of renal allografts (6). Hyperacute rejection has also been demonstrated in liver allografts in rhesus monkeys (7), pigs (8), and rats (9), where a single RT-1 mismatch may be sufficient to induce rejection (10). In rats there are rejector and nonrejector strain combinations. In the nonrejector, DA to PVG model grafts are accepted even in the presence of a positive cross-match (11), whereas in rejector combinations, such as LEW to BN, grafts given to presensitized animals may be rejected within days or hours of the operation (9, 10). Later experiments in rats suggested that the timing of donor-specific sensitization and the titer of antibody may be important factors in determining graft outcome (12). The pathology of rejected grafts in these experiments is consistent with an antibody-mediated immune rejection. Clinical observations. Most of the reported cases of hyperacute rejection of liver transplants are associated with ABO-incompatible grafts (13-15). Examination of failed ABO-incompatible grafts reveals enlarged hemorrhagic organs with widespread necrosis. The vascular picture is one of endothelial cell disruption, platelet margination, and varying degress of thrombosis. Biopsies and explanted grafts reveal complement, immunoglobulin, and fibrinogen deposition. Although a proportion of ABO-incompatible grafts do not develop hyperacute rejection (6, 15), the long-term outlook for those grafts that survive the early postoperative period is poor (16). Clinical evidence of hyperacute rejection of ABO-compatible liver allografts is limited to a few individual case reports (17-24). Most of these rare cases can be related to the presence of a strong positive donor-specific cross-match (17, 19, 20, 22-24), and in a number of cases there is evidence of immunoglobulin consumption by the graft (17, 19). However, the majority of cross-match-positive individuals receiving liver allografts do not develop hyperacute rejection. Indeed, the early analysis from Pittsburgh (25-27), which indicated that liver transplants may be performed in the presence of positive warm T cell cross-matches without the risk of hyperacute rejection, has been consistently verified in the numerous studies reviewed below. Combined kidney-liver transplantation. Although hyperacute rejection occurs in some rat strains (10), in the DA-to-PVG model, liver allografts may prevent rejection of other organ grafts from donors of the same strain but not third-party organs, even in the presence of preformed alloantibody (11). Similar protection has also been observed in clinical transplantation in cases requiring combined kidney and liver transplants where the recipient is cross-match positive (28). However, this procedure is not always successful, as evidenced by a report of two cases of combined kidney-liver transplantation aptly named “The Case of the Canary Kidneys” (29). In one of the two patients, the initial liver and a second liver transplanted 3 days later developed widespread necrosis, and although the patient was cross-match negative, C1q and IgM deposition were demonstrated in both the kidney and the liver. In the second patient, who was cross-match positive, there was massive necrosis of the liver, which remained intact, but the kidney never functioned. LONG-TERM OUTLOOK There are fewer studies of cross-matching and graft rejection/survival in liver transplantation than for other solid organ grafts (Tables 1 and 2). In addition, the debate over which cross-match technique is best and which antibodies should be considered in liver transplantation is less well developed than in renal transplantation. Overall, many of the published studies are poor, few studies define cross-match antibodies either by isotype and or specificity, and interpretation is made difficult by variable formats for reporting antibodies. In addition, the terms panel-reactive antibody (PRA*) and T-warm cross-match are unhelpful in the absence of Ig isotype and HLA or donor specificity. Comparisons between studies is further complicated by the use of different methods for antibody detection and clinical management policies, including primary immune suppression protocols, use of supplemental immune suppression, especially tacrolimus (FK506 rescue), antilymphocyte globulin, and OKT3. Acute and chronic graft rejection. Acute (early) rejection occurs in the majority of liver transplants and is believed to be mostly a cellular, primarily a T cell, phenomenon. Of the few recent studies that have investigated the role of cytotoxic cross-match antibodies in acute rejection, only one found a significant effect (30), whereas three other studies failed to detect any correlation with a positive cross-match (31-33). Although chronic rejection after liver transplantation occurs in only 8-9% of cases (34, 35), it remains one of the most common causes of graft loss in published series (36). Early studies suggested that neither pretransplant PRA nor donor-specific complement-dependent cytotoxic cross-match (CDC) was associated with chronic graft rejection or the so-called vanishing bile duct syndrome (VBDS) (25-27, 37, 38). Subsequently, there have been a number of conflicting reports; evidence for a relationship between VBDS and a positive cross-match was first published by the Mayo Clinic (39) and is supported by more recent reports from Pittsburgh (23, 40) and Houston (41), although recent data from transplant programs in Minnesota (42), the University of Virginia (31), Dallas (32), and King's College Hospital (33) do not support these observations. Although it is tempting to cite the sensitivity of the cross-match technique and the differences in the basic immune suppression regimens as reasons for the different findings, these alone may not be sufficient explanation. One feature of all of these recent reports is the low incidence of chronic rejection, and an alternative explanation may be that these studies have too few patients in the subgroups to give the necessary statistical power to the analysis; for example, in the study by Tayaka et al. (40), only 29 of 600 (4.8%) first grafts were lost to chronic rejection. Graft survival. Opinion is divided about the influence of cross-match antibodies on graft survival. The earliest large series reported no difference in 1- and 2-year graft survival in those with and without cross-match antibodies (26, 27). This is supported by recent reports from Toronto (30), the University of Virginia (31), and King's College Hospital (33). In contrast, recent data from Pittsburgh (23, 40, 43) suggest that T cell-positive cross-matches may have a detrimental effect on survival of first liver grafts. Similar findings have now been reported from Houston (41), UCLA (44), and a 10-year study from Dallas, where the effect of a T-warm cross-match was a striking 17% lower patient survival rate at 4 years (32). SELECTION AND SENSITIVITY OF CROSS-MATCH TECHNIQUE In the context of this Overview, the critical issues that need to be decided are: (1) which cross-match assay to perform (CDC with or without antihuman globulin, PRA, or, more recently, flow cytometry and enzyme-linked immunosorbent assay) and (2) which antibodies need to be considered in liver transplantation-IgG and/or IgM, T cell, B Cell, HLA class I or HLA class II, antiendothelial cell, etc. In the studies reviewed above, those that used the antihuman globulin technique detected the highest percentage of positive cross-matches (Table 1). 3However, with the exception of the data from the Mayo Clinic (39), the values obtained for CDC T cell cross-matches are relatively consistent, falling in a very limited range from 7% to 11%. Although flow cytometry is arguably the most sensitive technique available at present, it may give a higher rate of false-positive results, and standardization of the assay is difficult to achieve (45). In addition, whereas some preliminary data based on flow cytometric measurement of IgG binding to donor T cells suggest that there is a clear relationship between IgG flow positive cross-match and graft survival (44, 46) and also hyperacute rejection (47), other studies are less convincing. In at least one of these three studies (44), a similar, though less significant, effect was observed when patients were grouped according to the results of a standard IgG CDC cross-match (44). The use of flow cytometry to assess IgG binding to pooled Epstein-Barr virus cells may be superior to conventional PRA testing, but in our series the results do not correlate with graft outcome (33). By selecting flow cytometric assays for cross-matching, it is assumed that IgM antibodies are not clinically significant and that both complement-fixing and non-complement-fixing IgG antibodies have equal importance. In liver transplantation, this assumption may be premature. At least one report suggests that both IgG- and IgM-positive cross-match antibodies can adversely affect liver allografts (41), and although complement fixation is not believed to be a prerequisite for hyperacute rejection, the role of non-complement-fixing IgG antibodies in transplantation has not been widely investigated. Equally important in this debate is the question of antibody specificity and antibody titer. Not only do published reports often fail to discriminate between IgG and IgM antibodies, many substitute percentage PRA for antibody titer and few define the antibody specificities. Karuppan et al. (48) found that preformed antibodies directed at HLA antigens are associated with poor graft survival, whereas non-HLA antibodies had no overall effect on graft survival. An early study from King's College Hospital suggested that high-titer antibodies against HLA class I antigens produced after liver transplantation are associated with an increased risk of developing VBDS, whereas low-titer antibodies were not (37). Antibodies other than those directed at T cells may be important: B cell antibodies, which may represent weak HLA class I antibodies and HLA class II antibodies or may be directed against non-HLA antigens, can have a detrimental effect on liver grafts (30, 33), and in renal transplantation, IgM antibodies to epithelial cells, monocytes, and vascular endothelial cells may be important in chronic rejection (49-52). Future studies may also wish to consider the persistence of antibody after transplantation. Manez et al. (53) noted that cross-match-positive patients with persistent antibody after transplantation were more at risk of allograft failure than those who became cross-match antibody negative after transplantation. It may also be more fruitful to look for evidence of antibody-mediated graft damage, such as immunoglobulin deposition, which has been seen on biliary epithelium and vascular endothelium in biopsies of patients with chronic rejection (54), or to relate the morphology of biopsies in patients with IgG-positive cross-matches (55). MECHANISMS OF RESISTANCE TO ANTIBODY The data reviewed above suggest that liver transplants are partially resistant to antibody-mediated graft rejection. Thus, although antibodies may contribute to both hyperacute and chronic graft rejection, the majority of patients are unaffected by the presence of preformed antibodies against donor antigens. Resistance may be due to a combination of immunological and physiological factors. In practice, donor-specific antibodies may disappear from circulation at the time of operation (31). This may be due to immune absorption of antibody by the graft or formation of immune complexes with soluble HLA class I (sHLA-I) antigen shed from the liver (56, 57), or may occur as a result of dilution with transfused blood (58). Immune absorption may be enhanced by phagocytosis of the sHLA-I antibody complexes by the Kupffer cells (59, 60) and by regulation of the idiotype anti-idiotype network (61, 62). Soluble HLA may also be capable of inhibiting alloreactive cytotoxic lymphocytes (63) and could interact both with alloantibody and directly with the T cell receptor. The interaction between sHLA-1 and the T cell receptor would have great potential to induce T cell anergy, one of the possible mechanisms of achieving graft tolerance. In addition, poor platelet and complement function associated with end-stage liver disease may also aid resistance in the early posttransplant phase (64). Physiological factors cited for resistance to antibody-mediated liver allograft rejection include the liver mass, the location of the graft within the peritoneal cavity, and the presence of a dual afferent blood supply (6). Historically, the comparatively large volumes of replacement blood that were commonly used at operation may have had the dual effects of diluting circulating antibodies and depleting host lymphocytes (58), and some of the differences between the earlier reports and more recent studies may be explained by the considerably reduced surgical blood requirements. In conclusion, the role of the cross-match test in liver transplantation is uncertain. There is no doubt that, with the numbers of transplants performed increasing each year and the wide range of cross-match techniques currently available, we are now more able to assess the impact of cross-match antibodies. A better understanding of these antibodies and the potential of the liver to resist antibody-mediated graft damage is urgently needed. The identification of factors that influence graft outcome may in future provide the basis for the allocation of donor organs. Perhaps more importantly, understanding the mechanisms of resistance may offer hope to highly sensitized patients awaiting kidney transplantation.